TechDrawl normally focuses on matters of immediate relevance to entrepreneurs, but occasionally I like to visit some of the problem solving that is being conducted at levels far above the two-person-and-a-dog startup world.  The Austin Forum provides a great monthly venue for that.

On July 6, Dr. Dan Stanzione (a computational expert) and Dr. Matthew Vaughn (a biologist), both affiliated with the Texas Advanced Computing Center at the University of Texas at Austin, described the iPlant Collaborative, a large-scale National Science Foundation project applying high-end computing, data, and software resources to the “grand challenges” of plant biology.  As you will see, these challenges directly affect our lives.

Dr. Stanzione talked about why plants are important, and, yes it’s food.  Since 1975, food consumption in the US has almost doubled to 3800 calories from 2000.  We’re eating 4 1/2 meals per day, and our children's generation will reverse the trend of ever increasing life expectancy.  We would need 3-4 earth-sized planets of arable land to feed the population if everyone in the world starting eating like we do. 

In the pre-agricultural era 1 person could live off 6000 acres.  Now that number is 1/2 acre, and by 2050 there will only be 1/3 acre available per capita.

In the US diet, corn is the most important plant.  We’ve been able to increase production first with hybrids, then with biotechnology.  The Moore’s Law equivalent is that production has doubled about every 25 years since 1935, but as with Moore’s Law, we are reaching limits.

Corn cannot reproduce without human help.  The domestic corn we know bears no resemblance to corn found in the wild.  Yet domestication decreases genetic diversity and increases the risk of disease.  One virus can run amuck if all the corn comes from the same gene pool.

There are some societal trade-offs with respect to corn production.  Productivity has been flattening the last 15 years or so after the easiest yield gains have been taken.  Now there’s the question of renewable energy – the grain allocated annually for ethanol could feed 350 Million people!  And, commodity food price increases foster political instability and had much to do with the turmoil in the Arab world this spring.  The Fed’s QE2 program had a direct impact on corn price rises; it is being said that the Fed is proving more effective than all previous political and military efforts in toppling dictatorships.

Dr. Stanzione emphasized that the biology of corn crops is not all that well understood.  Cross breeding must be repeated periodically lest the gains be lost.  Enter then the world of genetic engineering, and even there the definition of a gene is not well defined.  There is a very simple plant, the Arabidopsis, that was described as the “white mice” of biology.  Even for that species, we understand only about 15% of its genes, and for rice or maize that number drops to about 1%.

Not that scientists aren’t working on this -- there are about 2300 new academic papers per day in this area.  There are more than 70 Million gene sequences that have been identified, and thus there is the need for computing power to derive practical insight from all this data.  IPlant is building the cyberinfrastructure for this purpose.

To describe the genetics of the plants on our planet would require building massive “phylogenetic trees,” yet 228 species would require more trees than there are atoms in the universe.  The state of the art that iPlant is trying to advance is a mixed combination of computational tools: labor intensive, fragile, hard to share, and hard to reproduce results.  Dr. Stanzione described one recent program that was estimated to require 1634 years to run (a very long time to watch that spinning pinwheel on your Mac), but that a few weeks of optimization using antique Fortran saved a millennium of computation!

Dr. Vaughn then talked about 3 basic topics that are examined in biological research:

• Genotype – the genetic structure of an organism, the double helix with its GCTA markers
• Environment – the effects of normal versus stressful environments
• Phenotype - the trait we care about - yield, nutrition, color, etc. When you buy the reddest tomatoes in the basket, you are making your choice based on the phenotype.  (Ask your grocer.)

He noted that there was a major advance in gene sequencing in 2005 and that biologists now need to command a computer cluster more so than spending time at a bench.  The massive new data means that theories can be tested without having to extrapolate from the lowly Arabidopsis.  Untangling gene networks leads to practical issues like reengineering rice to be water efficient like corn and avert the affects of water supply instability due to predicted climate changes.

Even robotics is coming into play. Phenotype measurement is still labor intensive, but robots are being built to collect and identify plant materials.

In summation, bioinformatics is changing rapidly, particularly with the boost from iPlant, and at the very least should keep us all fed with corn for the remainder of our lives.